Data Sharing Services on PhysioNet

By building a protected data archive using PhysioNet's services from the outset of your project, you can back up your data securely as you gather them. At the conclusion of your project, you can have a data set that is ready to be shared and that can stimulate and support further work on your research topic.

All data shared via PhysioNet must be de-identified (i.e., it must not be possible to associate data obtained from human subjects with any individual subjects). PhysioNet's free and open-source deid software may be useful for removing PHI from text files.

Once your data have been de-identified, they can be transmitted to a protected data archive that we will establish for you on a PhysioNet server (PhysioNet Works). Data can remain protected until a date chosen by you (typically at the conclusion of your project, or at a milestone such as publication of your major findings). At that time, our expectation is that your data will be transferred to PhysioBank, the publicly accessible data archives of PhysioNet.

PhysioNet Services:

• Researchers can upload data to protected storage using a web browser.
• The investigator may share access to the protected data archive with colleagues if he or she wishes to do so.
• PhysioNet provides daily incremental backup of all protected data archives at its primary location and monthly full backup at a second location. Data deleted before the most recent full backup are not recoverable. All disks used for storage (on-line or backup) are encrypted and kept in physically secure locations.
• Any type of research data may be stored, including data in spreadsheets, relational databases, or flat binary or text files.
• Data stored in PhysioBank-compatible formats may be viewed and processed using the PhysioBank ATM tools. These include tools that can be used to convert a wide variety of data formats into PhysioBank-compatible formats. The ATM toolkit may be particularly useful for investigators who need to develop expert annotations for their data.
• Up to 10 gigabytes of protected storage is available on a self-service trial basis. Data stored on this basis may be retrieved only from the same IP address from which they were uploaded, and they are permanently deleted after 10 days (no backup is preserved).
• Qualified researchers may apply for protected data archives to be established for their projects. These archives accomodate larger amounts of data, for longer periods.
• Archives sufficient to store up to 100 gigabytes are available without charge. Additional protected storage is available for a fee. There is no charge if the data are made available for immediate public release.
• Two hours of support are more than sufficient for most projects, and are available without charge. Additional support is available for a fee.
• PhysioNet Works is built using free and open source software and standard PC hardware. If you wish, you may create a fully functional mirror (copy) of the PhysioNet Works server infrastructure and of your protected data archive, which can be synchronized with the master PhysioNet Works server at daily (or less frequent) intervals. You are under no obligation to do so.
• Once the data have been released for posting in a public area of PhysioNet, there are no recurring charges.

Researchers who are allocated a protected data archive on PhysioNet Works agree that they and any colleagues who share it will not use it to store PHI, or any material that they do not have permission to copy, and that they will release their data for posting in a public area of the PhysioNet web site within a reasonable period. NIH grant recipients who agree to these terms are routinely allocated protected storage; we review requests from other researchers and grant allocations for projects that are in line with PhysioNet's goals as space permits.

PhysioBank: Public data archives

When your data are ready to be shared with other researchers, they can be submitted for our review prior to posting them in PhysioBank, the web-accessible public data archives of PhysioNet. See the Guidelines for contributors to PhysioBank for details of the criteria we use while reviewing contributed data sets.

Physionet Services:

• If your data have not previously been stored on PhysioNet Works:
  • Apply for a protected data archive, and indicate that the data are available for immediate review.
  • When your application is approved, you will receive instructions for uploading your de-identified data into your protected data archive.
  • When your upload is complete, let us know and we will begin our review.
• If your data are already stored on PhysioNet Works, we will begin our review on the agreed review date, or at any earlier date if you ask us to do so.
• Based on the documentation you have given us, we will prepare a home page for your data, and a news item to announce its availability, both of which you will have an opportunity to review before your data are posted in PhysioBank.
• Your data will remain in the protected data archive until the review is complete.
• Once your data have been made available to other researchers in PhysioBank, PhysioNet fields most questions from users unless you wish to handle them yourself.

Your grant application may require you to include a plan to make the data you collect in the course of your publicly funded research publicly available.

PhysioNet Services:

• PhysioNet provides a model data sharing plan that you can include in your grant application.
• If you wish to use PhysioNet services as part of your data sharing plan, we can review and provide guidance on it before you submit your grant application.
• If you propose and we accept a data sharing plan that makes use of PhysioNet services, we will provide a letter of support that can be included in your grant application. The letter will certify that if your project is funded, we will provide it with the PhysioNet resources needed under the terms of the proposed data sharing plan.